Nari Qwen3-TTS and Qwen3-ASR

Nari Labs offers optimized open-source models behind production-ready APIs for speech-to-text (ASR) and text-to-speech (TTS). Their Qwen3-ASR and Qwen3-TTS models rank top in Coval’s industry voice AI benchmarks for latency and accuracy, providing cost-effective voice AI solutions including public beta endpoints and options for inference and training support.